AI can write code now.
Hire the engineers who can build_
Standardized early technical signal at scale, backed by assessment science that holds up under scrutiny.
- Real engineering environments: 1,300+ work simulations across 80+ languages and frameworks in a full VS Code workspace
- AI posture set per assessment: enable or disable the in-environment AI Assistant, with every interaction captured as reviewable AI activity
- Defensible outcomes: every task validated by occupational psychologists, with adverse impact monitoring at every cut score
Why early technical signal is harder than ever
AI-generated submissions blur the signal
Candidates can produce working code without understanding it. Traditional screening cannot distinguish engineers who build from engineers who prompt. You need a way to verify what someone actually knows.
High-volume pipelines need structured evaluation
Ad hoc processes break at scale. When every interviewer runs their own format, signal quality varies across the team and every hire carries different evidence behind it.
Hiring decisions face increasing scrutiny
From the EU AI Act to adverse impact audits, regulators and candidates are asking harder questions about how hiring decisions are made. An undocumented process is a liability.
Generic tools miss what matters
Toy problems and algorithm puzzles measure recall, not engineering ability. They do not predict how someone debugs production code, reviews AI output, or works across a real codebase.
Configure, assess, decide
Screen combines real engineering environments, layered integrity signals, and documented methodology in a single assessment workflow.
Configure assessments that match real work
Build assessments in a full VS Code environment with terminal access, packages, and multi-file projects. Choose from 1,300+ work simulations mapped to the Codility Engineering Skills Model across 80+ languages and frameworks.
Enable or disable the AI Assistant per assessment. Every candidate gets the same environment and the same rules, and every AI interaction is captured as reviewable AI activity.
Candidates work the way your engineers work
Capture rich, reviewable evidence
Layered integrity signals track how the work was done: behavioral monitoring, similarity checks, paste detection, network IP checks, and AI Follow-Up Questions that verify whether a candidate understands the code they wrote. AI Follow-Up Questions and AI Candidate Feedback are available in account settings for every package, off by default until an admin turns them on.
Newer signals add Typing Pattern DetectionPreview, which flags line-by-line retyping from an external AI source, and Cheating apps detectionPreview through the Codility Desktop App, which surfaces AI helper tools that hide from screen sharing and routes the evidence to a human reviewer rather than auto-failing a candidate.
Effective time-on-task calculation and video proctoring are more accurate and tighter than before, so reviewers see a clearer picture of how a candidate actually spent the session.
Integrity Risk aggregates these into one reviewable level: None, Low, Moderate, or High. Reviewers see which signals contributed and drill into the evidence.
Signal on engineering ability, not just output
Assessment integrity
No issuesDecide with a defensible record
Structured scoring backed by documented assessment methodology. Every task reviewed by occupational psychologists. Adverse impact monitored at every cut score against EEOC guidelines.
When a Screen report with a VS Code task moves forward, the candidate’s solution now loads automatically into the follow-on Interview, so your interviewer starts from what the candidate already built. The full VS Code experience inside Screen itself remains in Preview.
The cApStAn linguistic audit confirmed 65% of tasks at or below B1 CEFR, designed for global fairness across non-native English speakers.
Decisions that hold up when questioned
Screen report · Summary
What other tools miss
Typical technical screening tools
Codility Screen
Screen is where it starts
Codility extends the same validated methodology from screening through interviews and into your existing workforce.
Interview
Structured technical interviews in a shared VS Code environment with sidecar services, whiteboard, and full transcript. Replace ad hoc technical interviews with a repeatable, evidence-backed process.
Skills Intelligence
Map and verify technical capability across the engineering org using the same validated methodology. Staff projects on proven skills, target development where gaps actually exist, and report AI readiness with evidence.
